There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks
Yes. You can check the open source code in sources\libs\nvdsinfer\nvdsinfer_context_impl.cpp
NvDsInferStatus
NvDsInferContextImpl::queueInputBatch(NvDsInferContextBatchInput &batchInput)
Because the data sent to tensorRT has already been normalized, even if you save it as an image, it may not be recognized.